Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Yevgeniya Litvinova
Feature Info – Improving the visualization
and usability of GIS background information
in the context of a mobile tourist application
FEATURE INFO – IMPROVING THE VISUALIZATION AND
USABILITY OF GIS BACKGROUND INFORMATION IN THE
CONTEXT OF A MOBILE TOURIST APPLICATION
Dissertation supervised by
Prof. Dr. Christian Kray
Prof. Dr. Pedro Cabral
Dr. Morin Ostkamp
February 2018
II
ACKNOWLEDGEMENTS
I would like to acknowledge my supervisors Prof. Dr. Christian Kray from the
Institute for Geoinformatics, University of Mūnster, and Dr. Morin Ostkamp from con
terra GmbH for guiding and assisting me with carrying out my research and writing
this thesis. Your critique has always been constructive and motivating and I enjoyed
working with and learning from you.
I would like to thank Marcel van Remmerden for many discussions where he
shared his experience that expanded my background knowledge of the user experience
domain.
A special word of thanks goes to Wayne, my partner. This thesis could not
have been realized without your support and encouragement. Finally, my family:
mom, dad, and my brother, for being the best family I could ever ask for and teaching
me to be inquisitive, persistent, and not afraid to seek help.
III
FEATURE INFO – IMPROVING THE VISUALIZATION AND
USABILITY OF GIS BACKGROUND INFORMATION IN THE
CONTEXT OF A MOBILE TOURIST APPLICATION
ABSTRACT
Feature info is an interactive map tool that allows viewing background information
about a map feature in response to a user action. Mobile devices come with a number
of limitations, e.g. small screen real estate and the variety of screen sizes, that can
affect map and feature info usability. Existing recommendations for feature info
design [1], [2], [3] focus mostly on content, i.e. effective communication of data,
leaving the “interaction design” aspect overlooked. This Master’s Thesis attempts at
improving the visualization and usability of feature info interfaces in the context of
mobile tourist applications and presents nine feature info design guidelines that
address specific usability problems. The problems were identified through heuristic
evaluation of five tourist applications and a user interview. Literature survey and two
usability experts provided ideas regarding how several of these problems can be
resolved. Three guidelines were evaluated in a lab-based usability test with twenty
participants. On average, assessed guidelines demonstrated a significant positive
effect on feature info usability by decreasing task completion time by 33% and
increasing task completion rate and System Usability Scale (SUS) score by 26% and
28% respectively. Proposed guidelines are not restricted by any specific use case and
can be applied to other application domains. Researchers and business practitioners
can use the guidelines as a reference in their daily work.
IV
KEYWORDS
Mobile Human-Computer Interaction
Usability
Usability evaluation
Interactive maps
UX/UI design
User interfaces
User-Centered Design
V
ACRONYMS
GIS – Geographic Information Systems
GI – Geographic Information
HCI – Human-Computer Interaction
POI – Point of Interest
PWA – Progressive Web Application
SUS – System Usability Scale
UCD – User-Centered Design
UI – User Interface
UX – User Experience
WebGIS – Web-based Geographic Information Systems
VI
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ....................................................................................... II
KEYWORDS ............................................................................................................ IV
ACRONYMS ............................................................................................................. V
TABLE OF CONTENTS .......................................................................................... VI
INDEX OF FIGURES ............................................................................................ VIII
INDEX OF TABLES ................................................................................................ IX
1. INTRODUCTION ............................................................................................... 1
1.1. What is Feature Info? ................................................................................... 1
1.2. Use Case: con terra App ............................................................................... 5
1.3. Motivation and Research Questions ............................................................. 7
2. RELATED WORK ............................................................................................. 8
2.1. Usability of WebGIS .................................................................................... 8
2.2. Usability of Mobile Maps ........................................................................... 10
2.3. Feature Info Status Quo .............................................................................. 11
3. METHODOLOGY ............................................................................................ 16
4. ANALYZING FEATURE INFO USABILITY ................................................ 20
4.1. Heuristic Evaluation ................................................................................... 20
4.2. User Interview ............................................................................................ 25
4.3. Questionnaire for Experts ........................................................................... 28
5. GUIDELINES FOR FEATURE INFO DESIGN ............................................. 31
6. EVALUATING GUIDELINES ........................................................................ 39
7. RESULTS AND DISCUSSION ....................................................................... 46
VII
8. CONCLUSION ................................................................................................. 53
REFERENCES .......................................................................................................... 54
APPENDIX ............................................................................................................... 61
A. List of usability problems detected during heuristic evaluation ................. 61
B. Level of detail and actions offered in feature info in evaluated mobile
applications ............................................................................................................ 64
C. User questionnaire ...................................................................................... 66
D. Expert questionnaire ................................................................................... 68
E. Tasks designed for prototype usability test ................................................ 70
F. Declaration of Academic Integrity ............................................................. 71
VIII
INDEX OF FIGURES
Figure 1. An example of a feature info [8]. ................................................................ 2
Figure 2. Feature info interface of Google Maps in 2005 [9]. ................................... 3
Figure 3. Feature Info windows in various mapping technologies. ........................... 4
Figure 4. Web application developed by con terra GmbH. ........................................ 6
Figure 5. User-centered design framework for WebGIS applications. ...................... 9
Figure 6. Common feature info layout types. ........................................................... 14
Figure 7. The number of attributes and actions shown in a single feature info by apps
studied in this thesis. ................................................................................................. 24
Figure 8. Google Maps uses icons to represent “Address” and “Open time”. ......... 33
Figure 9. Map (A) and Collect screen (B) interfaces of Collector for ArcGIS. ....... 35
Figure 10. Feature info window for two features. Currently displayed feature is
indicated by enumerator in the bottom right corner. ................................................. 36
Figure 11. City districts and tourist attractions shown in con terra app. .................. 37
Figure 12. Thumb reachability areas during one-handed interaction with a mobile
device [42]. ................................................................................................................ 38
Figure 13. Examples of prototyped feature info interfaces (A, Prototypes # 1 and 2.
B, Prototypes # 3 and 4. C, Prototypes # 5 and 6, D. Prototypes # 9 and 10). ......... 43
Figure 14. Task completion time and rate, and SUS scores for Prototype 1 (before)
and Prototype 2 (after) applying the guideline # 3. ................................................... 46
Figure 15. Task completion time and rate, and SUS scores for Prototype 3 (before)
and Prototype 4 (after) applying the guideline # 3. ................................................... 47
Figure 16. Task completion time and rate, and SUS scores for Prototype 6 (before)
and Prototype 5 (after) applying the guideline # 4. ................................................... 47
Figure 17. Task completion time and rate, and SUS scores for Prototype 8 (before)
and Prototype 7 (after) applying the guideline # 4. ................................................... 48
Figure 18. Task completion time and rate, and SUS scores for Prototype 9 (before)
and Prototype 10 (after) applying the guideline #8. .................................................. 49
Figure 19. Task completion time and rate, and SUS scores for Prototype 11 (before)
and Prototype 12 (after) applying the guideline #8. .................................................. 50
IX
INDEX OF TABLES
Table 1. Research methods used in this thesis .......................................................... 16
Table 2. Summary of measurement and analysis types used in this thesis .............. 16
Table 3. Problems found during heuristic evaluation of the apps ............................ 22
Table 4. Demographic profile of user interview participants ................................... 25
Table 5. Feature info design guidelines .................................................................... 31
Table 6. The overview of conditions placed to produce prototypes ......................... 40
Table 7. The SUS version used in the usability test ................................................. 44
Table 8. Results obtained in this thesis linked to the research questions ................. 51
1
1. INTRODUCTION
This section describes the problem that motivated the author to carry out this
research. The first subsection presents the definition of “feature info” and its role in
map applications. It also describes the challenges of mobile development and how
they may affect feature info design. As this thesis is use case driven, the second
subsection introduces the main application under study and highlights usability
defects in its feature info. The third subsection outlines the goal of this thesis and
proposes research questions that are aimed at achieving it.
1.1. What is Feature Info?
Ubiquitous accessibility to the Internet, personal computers, and smartphones
have impacted the development of geographic information systems (GIS) in the last
30 years. In addition, broadband Internet and cloud technologies have made possible
the shift from desktop to web, leading to the advent of Web-based Geographic
Information Systems (WebGIS). Initially designed for experts, today online map
services and applications have become a mainstay in our daily life. Today the average
person relies on map-based technologies to determine where they are, how to get
where they are going, how to track their personal health, and so much more. They link
people and devices all over the globe allowing them to share spatial data, information,
and knowledge instantaneously and cost-effectively [4].
With the beginning of “map revolution”, where interactive maps are taking
over their static ancestors, cartographic research has started focusing on interaction
design in pursuit of finding best solutions for map interaction. Making map services
and applications more user-friendly and accessible has shifted the focus of industry
and academia to usability. The term “usability” refers to a quality of a product
describing to what degree it can be used by a specific user group in order to achieve
its specific goals in a specific context with effectiveness, efficiency, and satisfaction
(adapted from ISO 9241-210:2010 [5]). If a system is developed following the
2
principles of usability and user-centered design1 (UCD), it benefits not only users, but
also its creators through reduction in production and maintenance costs as well as
increase in productivity, popularity, and sales [6].
This thesis sets out to explore the usability of feature info interfaces. Feature
info (Figure 1) is an interactive map tool that retrieves background information about
a single geographic entity and presents it in a dialogue window that overlays the map.
The dialogue window appears as a result of an action: click, hover over, or tap applied
to select a feature of interest shown on the map. The tool has different names across
different GIS software and services: “pop-up balloon”, “pop-up”, “info window”,
however, to stay consistent the term “feature info” will be used further on in this
thesis. Feature info is a core map interaction element of many proprietary and open
source Web mapping technologies [7].
Figure 1. An example of a feature info [8].
Feature info notion has remained unchallenged during the last several years. For
instance, Figure 2 demonstrates a feature info interface from Google Maps back in
2005. Below in Figure 3 displayed are feature infos of six currently popular Web
mapping technologies. As GIS applications have evolved, feature info has maintained
its “classic” look - a pop-up balloon with background information anchored to a map
feature.
1 User-centered design – “approach to systems design and development that aims to make interactive
systems more usable by focusing on the use of the system and applying human factors/ergonomics and
usability knowledge and techniques” [5].
3
Figure 2. Feature info interface of Google Maps in 2005 [9].
Because technologies mentioned in Figure 3 are frequently used as platforms for
dissemination of geographical information and knowledge, map applications based on
them oftentimes inherit the default “classic” feature info look, unless map creators
invest their time and effort into customizing not only the content, but also the visual
appearance.
(A) Map by Google Maps JavaScript API [10]. (B) Map by ArcGIS Online [11].
4
(C) Map by Leaflet [12]. (D) Map by CARTO [13].
(E) Map by Mapbox [14]. (F) Map by OpenLayers [15].
Figure 3. Feature Info windows in various mapping technologies.
The aforementioned companies encourage map creators to customize feature
info (e.g. [13], [16]) by styling it to match the map’s user interface (UI) design and by
displaying content to users according to their needs. For example, general users may
prefer to see multimedia files (photos/videos) about a feature to get to know it better,
while experts may be satisfied with the ability to view just a few numbers describing
feature’s attributes. In some applications, apart from only displaying background
information about a feature of interest, feature info also incorporates additional
functionality. For example, certain action buttons, a form that can be filled out to
submit additional information about the probed feature, or sophisticated tools that
show temporal dynamics of attributes of the probed feature. It is, then, necessary to
distinguish between working with the map and working with the feature info. Based
on that, feature info dialogue windows will be an auxiliary or main interface element
respectively. Finally, feature info functionality and visual appearance should not be
5
overlooked. Rather, approached by applying design thinking2 and integrating users
into the design process [17].
Nowadays mobile map applications are just as popular, if not more, as desktop
web map services [4]. They came to existence with the wide spread of mobile
technologies starting from 1990’s and have grown beyond mapping to offering
location-based services that are used by the majority of smartphone owners [18] . Such
services include local recommenders, navigation, fitness-tracker apps etc.
The usability of mobile applications depends on a number of factors:
limitations of mobile devices, such as small screen real estate, various screen
sizes, single-window interfaces, connectivity, fragmented sessions, and
touchscreen input;
a wide variety of available devices and their operating systems;
different ways of deployment (Web vs. native apps) [19].
As a result, mobile map applications need a separate design approach when it
comes to their development. Let us have a look at a specific example: open feature
info window blocks a certain part of the map. Having obstructed map view may cost
users losing understanding of the environment. Compromising between two windows
on a mobile screen will not be possible without trade-offs, although map view should
be the first priority [20]. Conclusively, further usability studies of feature info
interfaces in mobile map applications are required.
1.2. Use Case: con terra App
The main application under study is a Web map service3 developed by con
terra GmbH (“the con terra app”) as this thesis is a collaboration between the
University of Münster and the aforementioned company. The application is a
technology demonstrator and offers geographic data about Cologne, namely city
districts, boroughs, and precincts grouped into Basic Data layer4; libraries, museums,
and schools – grouped into Education and Culture layer; and tourist attractions,
2 Design-thinking - “design-specific cognitive activities that designers apply during the process of
designing” [60]. 3 https://www.conterra.de/en/produkte/con-terra-solutionplatform/mapapps/produktdetails 4 Layer – “A collection of entities of the same geometric type (dimensionality)” [4].
6
playgrounds- and sports areas, and places of event – grouped into Recreation layer.
The con terra app can be accessed through any browser on any device and once loaded
the Recreation layer is shown by default what makes it looks like a typical tourist
application. Examples of feature info windows are shown in Figure 4. They present
only simple attribute data and have easily noticeable usability problems. For instance:
the heading with its large margins leaves 20 pixels-wide space for scrolling and
viewing background information; only one row of the table can be viewed at a time,
and there is missing information.
(A) (B)
Figure 4. Web application developed by con terra GmbH accessed via
Samsung Galaxy S8 (A, feature info opened. B, feature info scrolled).
Feature info interface in the con terra application has a lot of room for
improvement. Further formal usability inspection is required to detect additional
problems. It will be useful to look at other mobile applications of similar nature to see
the broader spectrum of usability defects in feature infos.
7
1.3. Motivation and Research Questions
The goal of this thesis is to improve the visualization and usability of feature
info interfaces in the context of mobile tourist applications. It is motivated by the lack
of scientifically-derived recommendations on feature info interaction design; and the
need of considering the limitations of mobile devices and producing high-usability
feature info interfaces. To achieve this goal, the following research questions are
proposed:
1. What usability problems occur when users interact with feature info in
mobile tourist applications?
2. How solutions addressing these problems can be generalized into
usability guidelines for feature info design?
These research questions will be answered by applying the UCD approach, i.e.
understanding problems that occur in feature infos, specifying design guidelines to
address these problems, producing design solutions based on the guidelines, and
evaluating them. The results obtained in thesis are supposed to help usability
researchers better understand feature info interfaces and how users interact with it.
The remainder of this thesis is structured as follows: section two discusses the
usability of WebGIS and mobile maps as well as existing literature focusing on feature
info design; section three discusses research methods applied in this thesis; section
four provides detailed description of how feature info usability data was analyzed and
answers the first research question; section five introduces feature info design
guidelines and answers the second research question; section six describes in detail
the usability test conducted to evaluate the proposed guidelines; section seven
discusses the results of the evaluation and this thesis, the limitations, and future work;
and section eight concludes this thesis.
8
2. RELATED WORK
Originally, research in WebGIS usability was focused on desktop scenario,
whereas today it has been increasingly revolving around mobile technology. This
section follows this pattern and discusses, first, the usability principles of WebGIS
services and, second, mobile map applications. The third subsection provides an
overview of existing literature that supports feature info design. In addition, it also
establishes feature info design dimensions and describes what research methods can
be applied to evaluate the usability of feature info interfaces.
2.1. Usability of WebGIS
Usability of WebGIS is an emerging research field and recent publications
discuss establishing methodologies to measure it, conducting lab and field
experiments to benchmark systems and improve them.
In order to evaluate the performance of WebGIS, their specific particularities
must be taken into account [21]. Schobesberger, for instance, developed a framework
for UCD and evaluation of Web map-based systems [17]. The framework suggests
various usability research strategies that depend on resources available to researchers
at any given stage of system design (Figure 5). As UCD advocates for prioritizing
user in order to design better products, in the context of this thesis it was important to
learn more about potential users of tourist GIS applications. Chang and Caneday [22]
found that the behavior and perception of users who seek tourist information via
WebGIS depends on sex, age, and education. These user characteristics, therefore,
must be taken into consideration when initiating any usability research for such
applications. Another study by the same authors [23] showed that two factors –
usefulness and playfulness – drive users to interact with WebGIS when executing
tourism-related tasks. Of particular interest is playfulness because it is associated with
the presence of interactive tools, such as feature info, that allow to manipulate spatial
and attribute data.
9
Figure 5. User-centered design framework for WebGIS applications. Methods in
bold type correspond to a minimal required strategy. Adapted by permission from
Springer Customer Service Centre GmbH: Springer Nature, Modern Trends in
Cartography, Integrating User and Usability Research in Web-Mapping Application
Design, Schobesberger D., [COPYRIGHT] (2014).
10
Although WebGIS has its own domain-specific usability principles proposed
by Komarkova et al. [24], it is very common that researchers apply Nielsen’s
heuristics [25] to design and evaluate user interfaces. Common usability problems
identified in desktop WebGIS relate to violation of cartographic principles,
accessibility, input and search problems [26]; map visualization and map tools [27];
violations of usability principles such as learnability, flexibility, robustness, and
aesthetics [28]. Usability problems associated with map tools in the above mentioned
studies include issues with map legend, calculate distance tool, scale bars, missing
control to turn on/off map layers, panning and zooming. Problems related to feature
info were not described, however authors suggested it as new tool that could provide
benefits to users. Roth and MacEachren [29] described interaction strategies users
apply to interact with map primitives (zoom, pan, retrieve [feature info] etc.) of a web-
based geoapplication. Rapid usage of feature info was indicative of the user being lost,
confused, and unaware of the filter tool. On the other hand, a single purposeful
instance of using feature info supported users in executing their tasks, although
slightly increasing efficiency.
2.2. Usability of Mobile Maps
Similar as with WebGIS, usefulness, ease of use, and satisfaction drive users
to interact with mobile map applications [30], [31]. Usefulness can be described as
how the content of an app is relevant to users, ease of use – how easily that content
can be extracted, and satisfaction – how they feel about using the app. UCD stresses
that users must decide and prioritize what content they want to see in a mobile app.
However, it is developers who are in charge of ease of use and usability of mobile
applications that both are determined by the choice of a development platform.
The recent evolution of progressive web5 apps (PWA) made it possible for
PWAs to expand their capabilities and offer almost the same functionalities as native
apps do, e.g. using device’s hardware: GPS, proximity sensors, accelerometer,
camera, storage, and other exclusive features. Web applications can only be accessed
via a mobile browser that has its own controls and functionalities. An example
5 https://whatwebcando.today/
11
provided further in this thesis demonstrated that users’ behavior was influenced by
continuous usage of native apps. Performing these habitual actions in a web app
resulted in a usability error.
Usability principles for mobile map-based applications have been established
and validated recently by Kuparinen [20]. These heuristics, however, lack a
comparative study against general heuristics that would allow judging their
effectiveness. Little if any research is available on usability of mobile map-based
applications and especially interactive map tools. Some studies have explored how
users’ background influenced their ways of zooming and panning [32]; how multiple
modalities (speech and touch) make it easier to interact with a mobile GIS [33], and
how searching is carried out in a context-aware location-based service with an
adaptive interface [34].
2.3. Feature Info Status Quo
Steady growth of map-based applications produced commercially and
voluntarily has led to sophisticated innovations in feature info design that appear
sporadically in a small number of map products. These innovations have a potential
of wider application if scientific evidence of their efficiency is obtained and properly
documented in academic publications.
GIS companies publish videos and documentation on feature info design
within the capabilities of their desktop software, however there exist very little
recommendations for mobile platform. Skarlatidou [1] presented trust guidelines for
WebGIS interface design in five dimensions: graphic design, content design, structure
design, and functionality design. However, only some of these guidelines are also
applicable to mobile feature info interfaces. Muehlenhaus outlined general design
guidelines specifically for the feature info in interactive and mobile devices [2]. He
stated: “On maps that make heavy use of info windows [feature info], as much thought
should be put into their design as is put into the design of the mapped area itself”.
This thesis proposes four design dimensions that constitute the feature info
interface: interaction technique (how to open/close and interact with it), placement
(where on the screen it will appear), content (what user will see when it opens), and
12
layout (or styling, i.e. how the user interface will look like). Below are presented
literature findings that provide design recommendations for each corresponding
dimension.
Interaction technique
Tapping (a “mobile” alternative to mouse clicking) as a gesture corresponds
to “select” user action [35] and map users always need to select a map feature for its
feature info to open. As a result, tapping is the most common gesture to open a feature
info window. A general rule of thumb is that only one feature info can be open at a
time [36]. Once open, user should be able to close it easily. Not only mobile screens
limit the size of the map displayed, they also require objects on the map to be larger
to match the touch target size. When the feature info window is active, it may occupy
only a small portion of the map or the entire screen. Depending on the content, it is
also possible to perform interactive actions within the feature info, for example
swiping, tilting, or scrolling.
Placement
By supplying information on top of what is shown on the map, the system
increases user’s cognitive workload – an important aspect of usability [37]. Placing
feature info window next to the map feature will stress the connection between them.
Placing it at the top or bottom of the screen may yield different results depending on
the context of use. There are several open-source plugins, e.g. Snazzy Info Window6
or Leaflet Responsive Popup7, that support responsive placement. These and other
empirically implemented ideas should also be supported by theoretical frameworks.
Content
Providing relevant, accurate, and reliable content is important in order to
support users in executing their tasks. Unless textual content is easily understood by
user, it must be supplied with explanations and definitions for any kind of terminology
and abbreviations, and other notations [25]. For instance, users participated in the
study [38] commented that “ID” of the object presented in the feature info was not
6 https://github.com/atmist/snazzy-info-window 7 https://github.com/yafred/leaflet-responsive-popup
13
informative, and knowing object type would be more beneficial. If this system-related
information is not meaningful to users, there is no need in including it in the feature
info. Frequent content updates must be carried out to ensure that users are supplied
with up-to-date information, and to avoid having missing data or broken links. Non-
technical explanations should be provided also for error messages, along with the
information about alternative ways for user to achieve the desirable goal/result [1].
Hennig and Vogler outlined content recommendations to guide feature info
design in web apps for teenage users [3]. The total of 5 recommendations were
derived through participatory design and focused on content requirements (e.g. youth-
specific language, multimedia) and its visual representation (labeling and highlighting
hyperlinks, high contrast between letters and background, different text size to
differentiate between title, headings, and text). Another research [39] also supports
simple language and avoiding technical jargon when presenting natural hazard
information to non-expert users. These two examples suggest that context of use is
crucial to feature info, and to evaluate mobile tourism systems it is necessary to
understand user’s travel behavior [40].
In the proliferating era of interactive technologies and big data, researchers
raise the question of useful integration of rich content into mobile maps [41]. It is no
longer what to tell, but rather how and why, and whether what is told relates to users
and adds value into their interaction with mobile map technologies.
Layout
One way to design feature info layout is according to structure design
guidelines from Skarlatidou [1]:
1. “Group content in a logical manner (information should be easy to find)
2. Textual information on different pages [feature info may be scrollable/have
tabs] should be grouped effectively and should be relevant to the context.
3. Titles, headings and subheadings should be meaningful and should help with
skipping paragraphs”.
Among many examples of feature info available on the Internet, it is possible
to delineate three common layout types: cards, table, and narrative (Figure 6). ESRI
14
has similarly proposed three feature info “display options” [16] that appear similar to
those proposed. Cards is an example of a completely custom layout where information
and action buttons are all shown separately as cards on a surface. Table is the simplest
form to display content since it only includes attribute names and values. Narrative
could be either information from a single attribute field or just text that surrounds the
attribute value, e.g. “Current water temperature [VALUE]”. The choice between these
types of layout should be made according to user needs and preferences. As mentioned
earlier, custom styling of feature info is often regarded as a mandatory. It must go
hand in hand with responsive sizing, which is essential in mobile context because
users access maps from a wide variety of devices.
(A) Cards (B) Table (C) Narrative
Figure 6. Common feature info layout types.
Current usability state of a feature info interface can be evaluated for each
dimension individually or together using human-computer interaction (HCI) research
methods [42]: interviews, camera studies, customer feedback, surveys,
questionnaires, usability testing, heuristic evaluation etc. Specific methods applied in
this thesis will be discussed in the next section. Collecting qualitative and quantitative
data through lab and field testing allows understanding user’s workflow with it,
finding usability defects and measuring precisely the impact these defects produce,
e.g. lowering efficiency or effectiveness. It is very common that researchers mix
methods in order to achieve their research goals [43]. Although mobile web
applications differ from native ones in a number of aspects, in terms of feature info
visual design and content presentation both can be easily compared.
15
This overview of related work summarized usability principles of WebGIS
and mobile map applications, methods of evaluating the degree of compliance to these
principles, common usability problems associated with map applications and feature
info in particular. Content, as one of the feature info usability dimensions, is the most
extensively studied and described in academic literature, whereas interaction
technique, placement, and layout are generally subjects to experimenting and finding
optimal solutions for specific use cases. The rest of this thesis will be dedicated to
exploring these three dimensions and establishing specific design guidelines related
to them.
16
3. METHODOLOGY
Research methodology of this thesis was based on the framework for user-
centered design and evaluation of Web map-based applications (Figure 5) [17]. The
summary of research methods applied to answer each of the research questions is
presented in Table 1. Measurements and analyses carried out as part of the applied
methods are shown in Table 2. How these methods work together and why they were
a good choice is described below.
Table 1. Research methods used in this thesis
Research question # 1
(analyzing usability problems)
Research question #2
(creating design guidelines)
Heuristic
evaluation
X
User interview X
Literature review X
Questionnaire for
experts
X
Usability test X
Table 2. Summary of measurement and analysis types used in this thesis
Qualitative Quantitative
Measurements Questionnaires for users and
experts, comments and
feedback obtained in the
course of the user interview
The number of usability
problems found during
heuristic evaluation,
usability testing of prototypes:
task completion time and rate,
and SUS scores
Analysis Interpretations of
questionnaire responses,
comparative analysis of apps
reviewed in heuristic
evaluation
Comparative analysis of
heuristic evaluation results,
Statistical analysis of the
usability test results
17
The first research question, finding and analyzing feature info usability
problems, was to be answered by testing existing systems. A heuristic evaluation of
five mobile tourist applications was conducted by the author of this thesis along with
a semi-structured user interview where users evaluated one of the five applications -
the con terra app. The decision to mix two methods was made because heuristic
evaluation done by one evaluator allows finding only 20 to 51% of the usability
problems in an interface under study [44]. Thus, an interview with users was
conducted in order to find additional problems that could have been missed during the
heuristic evaluation and to explore how users perceive feature info.
Heuristic Evaluation
A heuristic usability evaluation is a fast way to understand the current usability
state of an interface by comparing it against a set of predefined usability principles or
“heuristics”. It is called “fast” because it does not necessarily require user
participation. The evaluation can be carried out by both usability specialists and non-
specialists, although the former tend to produce better results [45]. Some of the
heuristics for map-based services (Kuparinen [20], Komarkova et al. [24]) mentioned
in Related Work section were not applicable for feature info evaluation since they
were developed to assess an entire system and not just an individual interface. Hence,
generic Nielsen’s heuristic for user interface design [25] were chosen.
User Interview
Interviews is popular qualitative research method that is also used to evaluate
the usability of existing information systems [46]. In contrast to heuristic evaluation,
users take part in the interview and share their experience while interacting with the
system either by thinking aloud or filling out a post-interview questionnaire. During
the interview conducted in this research, users only interacted with the con terra app.
The main reason for this was that the app was not ready for deployment and could still
benefit from participatory design, i.e. user feedback would provide more details about
how feature info could be improved. Another goal of the interview was to understand
how people with cartographic/GIS background perceive feature info as opposed to
people without one.
18
The second research question, creating feature info design guidelines, was to
be answered by proposing recommendations that would help eliminate usability
problems identified while answering the first research question. While some problems
already had solutions available in academic literature, others lacked theoretical
backing. Then, it was decided to seek answers from best-practice examples by sending
a questionnaire to industry experts to get qualitative insights into how they approach
feature info design and development. Altogether, literature review and responses from
the questionnaire contributed to deriving feature info guideline concepts. Lastly, a lab-
based usability test was chosen to make conclusions regarding how guidelines
influence user’s efficiency, effectiveness, and satisfaction.
Literature Review
Literature review is an essential step in carrying out any research project [47].
The review of related work presented in the previous section showed that feature info
interaction design was an area where further research was needed. Once usability
problems were analyzed, another round of literature review was carried out this time
looking for specific solutions to specific problems. Supporting literature was also
found to back up expert opinions.
Questionnaire for Experts
Same as interviews, questionnaires are widely used for qualitative research of
information systems [46]. This method is moderately complex in its execution: first,
it requires creating a list of unique questions, second, recruiting participants, and
finally, analyzing obtained qualitative data that may come in different levels of detail.
Gable also reported that questionnaires may not provide deep insights into the studied
phenomena [48]. One particular challenge in the course of this thesis was designing
the questionnaire. It required a lot of modifications to ensure it was formatted properly
and questions were easy to understand.
19
Usability Testing
Engaging target users into usability testing is encouraged by UCD. By letting
them interact with prototypes, developers can identify early on if concepts work as
expected and user requirements are met. If concepts contain usability problems, then
prototypes require another iteration of design and testing. The goal of the usability
test in this thesis was to compare feature info interface prototypes before and after
applying the guidelines. To achieve this goal, methods of observation and a System
Usability Scale (SUS) questionnaire [49] were employed. Observations consisted of
tracking task completion time (i.e. efficiency) and rate (i.e. effectiveness). SUS score
was a measure of perceived ease of use. SUS is a post-study standardized
questionnaire that offers a number of advantages: objectivity, replicability,
quantification using mathematics and statistics, economical to use, the results are easy
to communicate, and, of course, scientific generalization [50]. In addition, the
questionnaire produces reliable results even with small sample sizes, does not require
any license fee, and the scores can be easily calculated by downloading a spreadsheet
[51].
Summing up, methods used in this thesis were selected based on the amount
of resources available, i.e. personal, human, financial, and temporal. Further detailed
explanations of how each method was applied are provided in the next sections.
20
4. ANALYZING FEATURE INFO USABILITY
This section is dedicated to answering the first research question posed in this
thesis - what usability problems occur when users interact with feature info in mobile
tourist applications? Research methods applied to answer this question are described
in depth together with the obtained results. The first subsection covers heuristic
evaluation. The second – user interview. Additionally, the section touches upon the
second research question and explains how some of the problems can be resolved.
This knowledge and how it can be acquired is described in subsection three dedicated
to questionnaire for experts.
4.1. Heuristic Evaluation
Participants. The evaluation was performed by one evaluator, the author of
this thesis, who was not a usability expert, however had several years of active
experience with mobile map applications.
Apparatus and Material. Heuristic evaluation of feature info interfaces was
carried out for five mobile map applications: con terra app, TripAdvisor, Foursquare,
Google Maps app, and mobile version of Google Maps website. The applications were
selected based on their: category – tourist app, popularity – with more than 10 million
downloads (it is hard to imagine that apps with bad usability would be popular), and
platform – native and Web-based. Three of the apps analyzed were native (Google
Maps, TripAdvisor, Foursquare) and two were web-based (mobile version of Google
Maps website and con terra app).
After selecting the applications, a list of tasks was defined to help the evaluator
get familiar with their feature info interfaces:
1. Find Cologne on the map and zoom in to view the city.
2. Select any point of interest (POI) on the map, open its feature info and
scrutinize background information and actions presented.
3. Repeat the previous step 20 times with other POIs randomly selected from the
map.
21
Nielsen’s heuristics for user interface design [25] were used to detect usability
problems and categorize them. The severity of problems was rated according to the
following scale [52]:
0 = I don't agree that this is a usability problem at all
1 = Cosmetic problem only: need not be fixed unless extra time is available on
project
2 = Minor usability problem: fixing this should be given low priority
3 = Major usability problem: important to fix, so should be given high priority
4 = Usability catastrophe: imperative to fix this before product can be released.
In addition to heuristic evaluation, feature info interfaces of the applications
were analyzed to obtain additional details:
interaction technique (e.g. tap or double-tap),
placement (top, bottom, or next to the feature),
percentage of the screen occupied when minimized,
available modes (minimized or maximized),
layout (table, cards, narrative),
content amount (level of detail, actions in maximized feature info),
context (temporal – e.g. “Open Now”, spatial – “0.5 km from current
location”) and help, as well as,
means of handling missing information (blank field, hidden field, or a
suggestion to contribute).
The first five items from this list above related to feature info design dimensions
introduced in Section 2.3.
Procedure. For each application, the evaluation took around two to three
hours. While carrying out the predefined tasks, the evaluator recorded feature info
usability problems and assigned each problem a severity score. After the evaluation
was completed, additional details were collected including required estimations, i.e.
percentage of screen occupied, amount of content offered. To identify the proportion
of the screen that was covered by feature info, the area in pixels occupied by feature
info was divided by the total area in pixels of the screen. Measurements were carried
out in GIMP 2.8.22 image manipulating program. To measure the amount of content
22
presented in feature info, level of detail and actions available were counted. This was
done only for maximized feature infos because minimized usually offered just feature
name, address, and rating.
Results. The summary of problems identified during heuristic evaluations is
shown in Table 3 (the numbers in each cell represent the amount of discovered issues
with regards to that specific heuristic).
Table 3. Problems found during heuristic evaluation of the apps
Con
terra
app
Mobile
Maps
website
TripAdvisor Google
Maps
App
Foursquare
1. Visibility of system
status. 1
2. Match between
system and the real
world.
1 1
3. User control and
freedom. 4 1 1
4. Consistency and
standards. 4 1
5. Error prevention. 1
6. Recognition rather
than recall. 1
7. Flexibility and
efficiency of use. 2 2
8. Aesthetic and
minimalist design. 3 2 2
9. Help users recognize,
diagnose, and
recover from errors
1 1 1
10. Help and
documentation
Total number of errors 17 1 8 3 1
Major or catastrophic
problems 16 1 4 1 1
23
The full description of problems can be found in Appendix A. The greatest
number of usability defects, especially major and catastrophic, was found in the con
terra app. This result was expected since the application is only a demo product.
Among the released apps, TripAdvisor had more errors than Foursquare or Google
Maps (both the app and mobile website). Such outcome can be partially described by
the fact, that TripAdvisor offered the widest variety of content (Figure 7, Appendix
B) in comparison to the rest of the evaluated apps, and it might have been cumbersome
for developers to keep everything in order.
Because heuristic evaluation was conducted only by one non-expert evaluator,
some usability problems detected might have been misclassified. It was not seen as a
problem, because the goal of the evaluation was only to identify the problems and
Nielsen’s heuristics were used just for reference. There is also a possibility that some
problems detected were false positive.
A general trend among the analyzed apps shows that it takes one tap on the
map feature to open its feature info window in minimized mode, i.e. located at the
bottom of the map and covering a certain part of the map. Percentage of the screen
occupied by feature infos in minimized mode ranged from 10 (Foursquare) to 30%
(con terra app).
In four out of five apps, maximized feature info is presented in fullscreen mode
as a separate window that slides by swiping from left to right (TripAdvisor,
Foursquare) or from bottom up (Google Maps app, mobile version of Google Maps
website). In Google Maps app, it is also possible to view feature info in full-screen
mode by double-tapping on a map feature. In the con terra app, by tapping the
“maximize” icon in the upper right corner of the feature info, it does not maximize
into a fullscreen window, but moves from the bottom of the map closer to the selected
feature. As a result, it covers up a significant amount of the map and prevents from
viewing other POIs unless closed. It is worth noting, that in Google Maps certain
features with limited temporal extent (e.g. road construction) only have a minimized
feature info that presents the most relevant information about a feature (less than five
lines of text). In applications with feature info available in fullscreen mode
24
(TripAdvisor, Foursquare, Google Maps app, and mobile version of Google Maps
website) all information about a feature can be viewed by scrolling down the dialogue
window. Google Maps app and Foursquare also offer shortcut tabs that help users
access faster certain sections of the scrollable feature info window. Depending on the
operational layer that is viewed on the map (e.g. attractions, hotels, restaurants), the
layout of a feature info interface may change. Thus, the apps can be commended for
customizing the layout according to the content displayed.
Content and actions offered in the feature info interfaces deserve a separate
paragraph. If one chose to sort the apps analyzed based on the amount of content
offered (Appendix B) from the largest to the smallest, then he or she would get the
following result (Figure 7).
Figure 7. The number of attributes and actions shown in a single feature info
by apps studied in this thesis.
Con terra app currently provides only basic attribute information about POIs
in the form of text. On the other hand, Google Maps app and website, Foursquare, and
TripAdvisor altogether let users to:
view infographics, browse videos and photo gallery of the feature;
read, translate, filter, and write reviews;
email, call, share, save to favorites, and navigate to the feature;
check availability, make reservations and place orders;
open other websites to access real-time information;
3
17
30
35 36
C O N T E R R A A P P G O O G L E M A P S W E B S I T E ( M O B I L E
V E R S I O N )
F O U R S Q U A R E G O O G L E M A P S A P P
T R I P A D V I S O R
25
know average time spent at the feature and its popular times;
view upcoming events at the POI, other places located within the POI, and
also suggested similar POIs;
submit any missing information, suggest edits, and ask questions.
Each feature info provides minimum background information about the
feature, but enough so to allow users make decisions without accessing any second-
party sources, for example the feature’s official website. In crowd-sourced apps
(Google Maps, TripAdvisor, Foursquare) features located around the city center tend
to have more background information supplied by users (reviews, scores etc.), than
those on the periphery. This may be explained by tourists visiting only the most
popular places (usually around the city center) and sharing their experiences
afterwards.
When installed for the first time, none of the apps provide any tutorials or tips
on how to work with the map and the feature info. Developers, perhaps, assume that
by now tapping on a feature to retrieve more information is a conventional rule and
potential users of the app are expected to possess that knowledge.
4.2. User Interview
Participants. According to [44], having more than 5 evaluators allows finding
at least 75% of usability problems. The total of 6 users (age group: 20-30) took part
in the study (user demographics is presented in Table 4). None of the participants had
worked with the con terra app before. The study was approved by the Ethics
Committee of the Institute for Geoinformatics, University of Münster. Recruiting
participants and carrying out the study took one week.
Table 4. Demographic profile of user interview participants
Male Female
With GIS background 2 1
Without GIS background 1 2
Total 3 3
26
Apparatus and Material. Three tasks similar to those used during heuristic
evaluation were created to let users get acquainted with the con terra app:
1. Find Cologne on the map and zoom in to view the entire city.
Note: When the application is started, you will see markers on the map that represent
attractions and venues.
These markers altogether have been grouped into “Recreation” category. In addition
to this category, there are two others - Basic Data that shows city districts, boroughs,
and precincts; and Education and Culture that shows libraries, museums, schools.
Choose the category you want to view by clicking on the icon at the bottom
panel of the app.
2. Select any point of interest on the map (POI) and try to get the following
background information about it:
a) Name
b) Type
c) Address
3. Repeat the previous step 10 times with other POIs randomly selected on the
map.
While completing the tasks, users were requested to share their experience by
thinking out loud. Upon finishing, they were also requested to fill out a questionnaire
(Appendix C) that consisted of twelve open-ended questions designed by the author
of this thesis. Questions aimed at achieving the goals of this interview: recording
usability problems in the feature info of the con terra app, finding how users perceived
the feature info, and finding what users were missing in the feature info. Pilot
interview was conducted with one person, after which some questions were modified
for the sake of clarity.
Procedure. The interviews were held in a lab-based setting. At the start,
participants (or users) were requested to sign the informed consent. Then, they were
invited to get acquainted with the con terra app by performing the tasks. While users
were performing the tasks they were requested to think aloud and their comments
were recorded by the facilitator. After finishing the tasks, the participants were invited
to fill out the questionnaire. The responses have been anonymized and processed in
27
aggregate. Finally, usability problems detected during heuristic evaluation and user
interview were matched to eliminate duplicates.
Results. Users felt like feature info could benefit them by providing more
information about a single geographic entity, allowing to manipulate that information
through actions, and serving as an access point to new screens and interfaces that
would show even a deeper level of detail about specific aspects of interest. All users
indicated that they would like to view background information in an interactive way
because it could give better first impression of place and let them execute certain
actions right in the feature info (e.g. without having to copy the phone number and
paste in the dial screen). One user stated that he or she would like to customize feature
info to see only preferred content. Users were indifferent towards the layout of the
feature info and could not tell how recent the provided information was. They needed
the map often and five users shared that having minimized feature info still kept their
focus on the map, whereas fullscreen feature info would let them focus only on the
background information.
Four users liked the design of the feature info interface and especially that
external links were provided. However, there were more comments about particular
things they disliked, for instance the feature info was uncomfortable to use (icons were
difficult to select, scrolling area was too small, text font was too small); it did not
provide enough information (note: the app is only a demo); some did not understand
attribute names, and also they could not see the connection between the map feature
and its feature info. Participants also expressed what background information they
were missing: opening hours, pictures, distance from current location, prices,
popularity, contextual information (e.g. distance from current location or current
events at the venue).
The main result was recording six additional usability problems missed during
heuristic evaluation:
1. font size was too small;
2. the UI was difficult to understand;
28
3. “maximize” and “close” controls were small and located too closely to each
other – participants had problems using them;
4. system failed to respond when a feature was tapped on, however after zooming
in, it responded with opening a feature info;
5. missing information made table look incomplete; and, finally,
6. there was no visual cue to hint users that the feature info was scrollable.
4.3. Questionnaire for Experts
Participants. To find experts, LinkedIn8 was queried for specialists in UX,
front-end development, and usability, currently working in companies of mapping /
GIS industry, such as ESRI, Mapbox, CARTO, HERE, TripAdvisor, and Trivago, and
involved in the development of mobile apps. Out of eight prospect participants that
agreed to participate, only two submitted their responses. Recruiting participants and
collecting responses took three weeks. The study was approved by the Ethics
Committee of the Institute for Geoinformatics, University of Münster.
Apparatus and Material. The questionnaire (Appendix D) consisted of sixteen
open-ended questions that covered specific aspects related to feature info, such as
visual design, content, customization, and conducted usability studies. Questions were
designed in the way to obtain best-practice examples from mobile map application
and feature info development.
Procedure. After experts expressed their interest in participating, they were
provided with a consent form together with a questionnaire document, with a
submission deadline of two weeks. Once the responses were collected, they were
anonymized and processed in aggregate.
Results. Both respondents indicated that they approach feature info
development following user-centered design. One interviewee stated that “If the map
isn’t going to be critical for the user to get to the end goal, I make it a smaller feature
in the application, however if it is, I make it larger”. Visual design of feature infos in
different products usually remains the same and agrees with design guidelines
8 https://www.linkedin.com/
29
established by their respective companies. Feature info layout is chosen according to
the target audience, e.g. presenting background information as a table makes content
easy to scan and understand. However, a table may look too mechanical when it shows
“Walking duration: 1 minute 47 seconds” instead of “2-minute walk”.
With regards to feature info placement, if it is a feature users are expected to
use frequently, it is critical to make sure it is placed in a location that not only the user
can find, but can access quickly if need be. While one respondent suggested
experimenting in identifying optimal feature info shape, its dimensions and position
through user testing, another suggested trying to maximize available screen real estate,
prioritizing map view over feature info view, and taking into account reachability.
Any type of content can be presented in feature info as long as it is relevant to
the map feature, recent, and benefits the user. Providing dynamic context is important
because the feature info should be assisting users in real-time decision making. Also,
developers may want their app to be a “one-stop shop” providing the maximum
necessary information for users in order to support them in executing their tasks. Both
respondents also indicated that when designing feature info windows for mobile map
products, they only customize content presented in the feature info. One respondent
stated: “I haven’t come across a use case for personalizing the look of the feature info,
but of course there may be scenarios where this is needed”. Another argued that
“Giving users the ability to customize the look/content causes the user to potentially
lose focus of why they were there in the first place, which could cause a higher bounce
rate”. Suggested use cases where feature info customization might be relevant were
saving/bookmarking places and adding personal notes or comments, customizing
layout to access the most frequent actions/information, adding missing information,
i.e. flagging or suggesting an update.
Both experts indicated feature info discoverability as a main usability problem
many novice users face. Therefore, they suggested that having a feature info tutorial
or a walk-through during onboarding would benefit users.
Only two out of eight experts contacted, submitted their responses to the
questionnaire. This might have inhibited the author of this thesis from getting a deeper
30
understanding of how industry experts approach feature info design. In addition,
having more best-practice examples would allow producing even more specific
guidelines.
The results of the heuristic evaluation and user interview answered the first
research question and presented thirty-four unique usability problems in feature info
interfaces. The majority of those problems were the result of poor interaction design.
For some usability problems described in this section it was challenging to find
complete solutions in academic literature. Contacted usability experts shared their
opinions regarding how these and other problems they experienced could be resolved.
31
5. GUIDELINES FOR FEATURE INFO DESIGN
This section provides an answer to the second research question, i.e. how
solutions addressing these problems can be generalized into usability guidelines for
feature info design? It introduces the guidelines and explains them on a conceptual
level.
Feature info design guidelines that target specific usability defects are
presented in Table 5. Seven out of nine guidelines cover the “Design” aspect that is
related to usability dimensions (interaction technique, placement, and layout)
established in Section 2.3. The other two guidelines focus on “Content” and
“Awareness” (one for each category respectively). Guidelines:
# 2, #3, #5 were derived based on expert feedback;
#1 and #6 were derived based on existing solutions presented in studied
applications – Google Maps app, TripAdvisor, Foursquare;
#4, #7, #8, #9 were derived based on a combination of sources: expert
feedback or existing solutions further supported by research literature.
How each guideline was derived and what usability problem it addressed is further
elaborated below.
Table 5. Feature info design guidelines
Content
1. Provide context, e.g. measurement units for numerical values or relative rank of the
feature based on an attribute (e.g. area size or popularity), to help user understand
the relationship between the currently viewed feature and the rest of those presented
on the map.
Design
2. Choose feature info layout according to target audience. Users must decide and
prioritize what type of content they want to have in the feature info and how it should
be presented.
32
Table 5. Feature info design guidelines Continued
3. Avoid cramming more content by making minimized feature info window scrollable
and switch to maximized/full-screen window instead. The size of feature info should
be primarily guided by the amount of content it offers. Shape and dimensions of the
minimized feature info window must be chosen in a way that prioritizes map view.
4. When designing feature info interface, the control to return back to the map should
be clearly visible and touch-friendly.
5. Choose maximized feature info window to present interactive content because
interactive gestures - swipe, pan, or tap, may contradict those of the map.
6. To avoid selecting multiple features accidentally, cluster point features at small
scales and show them individually when zooming in; for polygon and line features,
make only labels selectable.
7. To ensure natural interaction, place minimized feature info window at the bottom
of screen.
8. Find a way to show the connection between the map and the feature info window
by bringing into focus the feature that was selected (e.g. change marker color / size).
Awareness
9. Feature info as a map tool struggles with the problem of discoverability. Provide a
tutorial or a walk-through during onboarding to let users know about it.
Guideline # 1: Providing measurement units and relative ranks. On a
mobile device, due to limited screen real estate it may be cumbersome to fit in the
map legend that displays maximum and minimum values and a measurement unit of
a layer currently displayed. When feature info is open, user will see the value of an
attribute at current feature without knowing the measurement unit or relative rank of
the attribute at this feature in comparison to other features (e.g. colder/bigger/more
populous than others). In tourist apps that were reviewed, e.g. features were always
ranked based on the ratings they receive from the people who visited them. It could
be a good practice to provide measurement units and a relative scale position of a
value characterizing a certain attribute in feature infos.
33
Guideline # 2: Choosing feature info layout. Six out of eight user interview
participants agreed that presenting background information formatted as a table makes
content easy to scan and understand. Especially when it comes to comparing
background information between several features. Knowing the target audience and
the type of content that will be displayed can definitely benefit the design process of
the feature info layout. When users need information urgently, having to process
narrative or long text may cause frustration, make user lose track and turn to other
sources for desired information. For advanced users of an app/system, table or cards
will be the fastest and easiest way to obtain information, often because they memorize
the names, measurement units, and order of the attributes in the feature info thanks to
their daily exposure to the map product they utilize. In other scenarios users might
prefer information presented in a natural conversational flow as if “offered by
Wikipedia” (quoted from one of the interview responses).
Ensuring that an attribute’s value can be fetched easily on demand, hence,
becomes the main task for the UX researcher. One way to make table look “friendlier”
is by experimenting with its visual design. Eliminating borders or making them subtler
may help drive more focus to the content. Certain terms used to describe attributes
may seem ambiguous or confusing, and providing context (text that surrounds the
attribute value, e.g. “You can fly your drone at [WIND SPEED VALUE] m/s”) will
make the interaction between users and the map app seem more natural. Another
option would be to use icons instead of words for certain commonly understandable
attributes (see Figure 8 below). By doing so, designers manage to achieve two goals
at the same time: keeping the design clean and intuitive and also saving screen real
estate. Icons created for any other attributes are suggested to undergo user testing to
ensure that they communicate the right information.
Figure 8. Google Maps uses icons to represent “Address” and “Open time”.
34
Guideline # 3: Choosing feature info window size. Another important aspect
is identifying the amount of content that will be presented and the need for a full-
screen feature info window. As one expert stated: “having a miniature feature info
window allows presenting the most relevant background information about a
geographic feature while still keeping the user’s focus on the map”.
According to Material Design, feature info definition is somewhat similar to
the “card”– “a sheet of material that serves as an entry point to more detailed
information” [53]. Material Design does not encourage making the card scrollable.
During user interview this rule was empirically proven by all users who participated
in the study and found it extremely difficult to scroll in a narrow space of the feature
info of the con terra app (users with “fat fingers” were frustrated the most). Moreover,
heuristic evaluation of the apps showed that none of the apps except the con terra app
allowed scrolling through minimized feature info window. If the amount of content
that needs to be presented in the minimized feature info window does not fully fit into
the window’s boundaries, instead of making it scrollable, expand it into a full-screen
window that can be scrolled internally. Vice versa, there is no need for full-screen
feature info window if no extra information is provided. To help users understand that
the feature info can be expanded into a full-screen window, a number of visual cues
can be implemented, such as animation (sliding in from the bottom would suggest that
the window is swipe-able); cropping content, or simply showing a “more info” button.
Alternatively, shortcut tabs can be offered to help users access certain sections of the
scrollable dialogue window faster, like in Google Maps or Foursquare.
Guideline # 4: Providing visible controls to close feature info. While
observing participants interacting with the con terra app during the user interview,
several of them had a problem closing the feature info. The reason for this could be
the fact that users of Android native apps were accustomed to using the system’s
“Back” button to exit an unwanted state. Applying the same action in a browser while
using a Web app, instead, returned users to a previously visited website. Thus, another
usability catastrophe was witnessed. When designing feature info windows, the
control to close them should be clearly visible and touch-friendly [2].
35
Guideline # 5: Choosing maximized feature info for interactive content.
All participants of the user interview wanted to see interactive content in the feature
info. When offering such content, designers should consider making it fullscreen,
because otherwise interactive gestures such as swipe/pan/tap may conflict with those
of the map. This practice is widely followed by apps evaluated during the heuristic
evaluation and also other mobile map apps (see Collector for ArcGIS example with a
form in Figure 9 below):
(A) (B)
Figure 9. Map (A) and Collect screen (B) interfaces of Collector for ArcGIS.
Guideline # 6: Making map features easy to select. Zoom level plays an
important role when users interact with the map and feature info. When multiple
features were located nearby (Figure 10), the app did not cluster feature markers,
making the map look cluttered.
Since the tap target size is usually considered to be 48 x 48 pixels, multiple
map features can fit into this square. Tapping on a feature marker at a small scale
sometimes opened feature info for another than selected map feature. Moreover, four
36
out of six users had problems with opening the feature info, i.e. system failed to
respond upon tap until they zoomed in to a larger scale. One way to tackle all these
issues would be to reduce marker overload by clustering and showing one cluster
marker with a number of features that belong to it. Tapping on a cluster marker should
bring map to a certain scale where each individual marker can be selected to view its
feature info separately.
Figure 10. Feature info window for two features. Currently displayed feature
is indicated by enumerator in the bottom right corner.
Finally, if two layers were displayed on the map at the same time (Figure 11),
e.g. a polygon and a point layer, and it happened so that a point was located within a
polygon, two users faced a situation when upon tapping the point, a feature info for
the polygon was shown instead and users were confused and unable to execute the
assigned task. This particular problem can be solved by placing a name label within
each polygon and making only the label clickable. Labels, of course, should be visible
only at a certain zoom level. In such way, even if there are point features within the
polygon, they can be selected individually to open their feature infos.
37
Figure 11. City districts and tourist attractions shown in con terra app.
Guideline # 7: Placing feature info to ensure natural interaction. Placing
feature info next to the selected feature can block a substantial area in the middle of
the map. User interview participants responded negatively to this while they were
using the con terra app. In general, users of mobile devices prefer to hold them in one
hand [54] and use that hand’s thumb for interaction. Thumbs can only reach so far and
limited thumb reachability narrows down the functional area of the screen (shown in
green in Figure 12). Keeping in mind these two ideas and also expert feedback, place
minimized feature info window at the bottom of screen to prevent it from blocking
the map and allow it to be reached easily.
38
Figure 12. Thumb reachability areas during one-handed interaction
with a mobile device [42].
Guideline # 8: Establishing mental connection between the map feature
and its feature info. If feature info is shown at the bottom of the screen, it is important
to provide visual distinction of the selected feature from the rest of those displayed on
the map, as done by TripAdvisor, Foursquare, and Google Maps. While listing the
aspects users liked the least in the con terra app, two users indicated that features were
difficult to distinguish between, and after certain amount of time they lost the track
which feature the currently displayed feature info window belongs to. Therefore, it is
highly important to show the link between the feature and its feature info window by
e.g. changing the color of the marker selected, dropping the shade, or increasing its
size (as suggested by users participated in the interview and [2]).
Guideline # 9: Supporting feature info discoverability. Both experts
indicated feature info discoverability as a main usability problem many novice users
face when working with mobile map apps. This claim was further supported by
Andrienko et al. [55] and by the fact that none of the tourist apps analyzed during the
heuristic evaluation provided any tips or hints on how to use feature info. Experts
commented that having a tutorial or a walk-through during onboarding might be
beneficial to users. While onboarding is quite common for native apps, product
developers must think carefully of how to implement it into Web apps. If every time
the link is clicked, a map opens up with an onboarding popup, users may get frustrated
with having to get rid of it. One solution could be having a “help” button on the map
that could provide tips for working with that map and its interactive tools.
39
6. EVALUATING GUIDELINES
How the proposed guidelines can affect a feature info interface was decided to
assess through a lab-based usability test. Due to the time constraint placed by the
deadlines of this thesis, only three out of nine guidelines were chosen for testing. In
addition, recruiting large sample of participants in short time appeared challenging,
therefore the test was designed to be within-group. The choice of guidelines was based
on the severity of problems that they were supposed to solve, and also on the fact that
the majority of users clearly expressed their frustration when they faced these
problems. The following hypotheses were proposed:
Hypothesis 1 - based on Guideline # 3: Maximized feature info design will
perform better than minimized window when serving user rich content:
Sub-hypothesis a: Long text
Sub-hypothesis b: Forms
Hypothesis 2 - based on Guideline #4: Visible control to close feature info
window will perform better than invisible.
Hypothesis 3 - based on Guideline # 8: The presence of visual feedback will
allow establishing and maintaining the mental connection between the selected feature
and its feature info window.
The effect that guidelines produce was assessed through performance and perception
measures: task completion time and rate, and SUS score.
Participants. Twenty participants (aged 20-30, 10 males and 10 females, 10
with GIS background) were recruited to take part in the usability test. Participants had
to have knowledge and experience of using mobile tourist applications. The study was
approved by the Ethics Committee of the Institute for Geoinformatics, University of
Münster. Recruiting participants and carrying out the test took two weeks.
Apparatus and Material. Selected guidelines were applied to redesign feature
infos of two applications: the con terra app and Google Maps. The former was chosen
because it was the source of usability problems that provided the basis for the
40
guidelines under study. Whereas the latter was chosen because of its popularity [56],
[57]. Having used it on a daily basis, users acquired a certain behavior where they
became accustomed to certain controls and layouts. It was of interest to observe how
users’ behavior would change after induced interface changes. The total of twelve
prototypes (Table 6) were created to test the hypotheses.
Table 6. The overview of conditions placed to produce prototypes
Prototype
#
App Stimuli
Hypothesis 1a
1. con terra app Small feature info
2. con terra app Fullscreen feature info
Hypothesis 1b
3. Google Maps app Small feature info
4. Google Maps app Fullscreen feature info
Hypothesis 2
5. con terra app Visible control to close small feature info
(“X” button)
6. con terra app Invisible control to close small feature info
(tapping outside feature info)
7. Google Maps
app
Visible control to close fullscreen feature info
(“Back to the map” button)
8. Google Maps app Invisible control to close fullscreen feature info
(use Android’s Back button)
Hypothesis 3
9. con terra app Presence of visual feedback for small feature info
10. con terra app Absence of visual feedback for small feature info
11. Google Maps app Presence of visual feedback for fullscreen feature
info
12. Google Maps app Absence of visual feedback for fullscreen feature
info
41
The prototypes were designed in Balsamiq Mockups9 v.3.5.15. Static
prototypes exported in PNG-format were transferred to a Samsung Galaxy S8 mobile
device and uploaded into POP 2.0 - Prototyping on Paper10 app to create interactive
prototypes by linking images with one another. Examples of prototypes are shown in
Figure 13.
Twelve tasks typical for tourist map applications (e.g. finding phone number,
ordering cinema tickets, browsing restaurants) were created for the test (Appendix E).
All tasks were designed to be independent from one another.
Observed data were recorded by the facilitator (the author of this thesis). Task
completion time was tracked using a simple stopwatch from the moment user started
interacting with the prototype until reaching a predefined goal. Completion was
recorded using binary digits: 1 - for success and 0 - for failure. Although SUS is a
post-study questionnaire, in this usability test, it was used to measure perceived
usability after each task. The reason for this, was that technically each prototype
represented a different [feature info] system. Original SUS questions were modified
by substituting “system” for “feature info” (Table 7). According to Lewis and Sauro
[58], such substitution does not affect the resulting scores.
Google Forms11 were used to collect and organize SUS data. Later, the scores
were calculated using the spreadsheet [51]. A pilot test was conducted with one
participant, after which some tasks were modified for clarity.
9 https://balsamiq.com/ 10 https://marvelapp.com/pop/ 11 https://www.google.com/forms/about/
42
(A)
(B)
43
(C)
(D)
Figure 13. Examples of prototyped feature info interfaces
(A, Prototypes # 1 and 2. B, Prototypes # 3 and 4.
C, Prototypes # 5 and 6, D. Prototypes # 9 and 10).
44
Table 7. The SUS version used in the usability test
Strongly Strongly
disagree agree
1 2 3 4 5
1. I think that I would like to use this feature info.
2. I found the feature info unnecessarily complex.
3. I thought the feature info was easy to use.
4. I think that I would need the support of a technical
person to be able to use this feature info.
5. I found the various functions in the feature info were
well integrated.
6. I thought there was too much inconsistency in this
feature info.
7. I would imagine that most people would learn to use
this feature info very quickly.
8. I found the feature info very cumbersome to use.
9. I felt very confident using the feature info.
10. I needed to learn a lot of things before I could get going
with this feature info.
Procedure. Participants were invited to participate in the test individually. The
test was conducted in a lab-based setting with one facilitator. Prior to starting the test,
an informed consent was collected from each participant. Following the explanation
of the test procedure, users were presented a practice prototype to get acquainted with
the test. Once they were ready to start, printed tasks were randomly mixed and users
drew one task at a time. After completing the task, users shared their experience by
filling out the SUS questionnaire. The executed task was then removed from the pool
to avoid having some tasks not performed. Before each experiment, all tasks were put
back together and shuffled to randomize their order and avoid learning effect. Average
test duration was 35 minutes.
Data analysis. Raw performance and perception data were reported using box
plots. Shapiro-Wilk Normality Test was performed to verify if collected data was
normally distributed (p < 0.05). Non-parametric Wilcoxon Signed-rank test was
performed to identify the effects of guidelines on task completion time and SUS
scores. This test was chosen because the general assumptions for parametric tests were
45
not met: some populations were not distributed normally and some were measured
through ordinal variables (SUS scores were based on Likert-type scale). Since task
completion data was categorical (1 or 0), McNemar’s test, a paired version of Chi-
square test, was applied to determine how guidelines affect completion rate. Spearman
correlation test was performed to identify the effect of GIS background on
participant’s performance and perception. Data analysis was performed in R Studio
Version 1.0.143. For Wilcoxon test, an additional package ‘coin’ [59] was installed
to address the problem of calculating the exact p-values because of ties.
46
7. RESULTS AND DISCUSSION
Prototypes 1 and 2. Fullscreen feature info demonstrated higher usability
when users had to work with long text. On average, applying the guideline reduced
task completion time by 15% and increased SUS score by 20% (Figure 14). Task
completion rate also decreased by 5% only because 1 out of 20 participants failed to
complete Task 2. The reason for failing was that the participant misunderstood the
task. The guideline # 3 was found to have insignificant effect on task completion time
(W = 143, Z = 1.42, p > 0.1, r = 0.22), but significant effect on task completion rate
( (1, N = 40) = 17.05, p < 0.001) and SUS score (W = 14.5, Z = -2.82, p < 0.005, r
= 0.45). Obtained results supported Hypothesis 1a.
Figure 14. Task completion time and rate, and SUS scores for Prototype 1
(before) and Prototype 2 (after) applying the guideline # 3.
Prototypes 3 and 4. Fullscreen feature info demonstrated higher usability also
when users had to work with a form. On average, applying the guideline reduced task
completion time by 32%, and increased task completion rate and SUS score by 10 and
38% respectively (Figure 15). The guideline # 3 was found to have significant effect
on all three: task completion time (W = 195, Z = 3.35, p < 0.001, r = 0.53), task
completion rate ( (1, N = 40) = 13.14, p < 0.001), and SUS score (W = 6, Z = -3.64,
p < 0.001, r = 0.58). Obtained results supported Hypothesis 1b.
47
Figure 15. Task completion time and rate, and SUS scores for Prototype 3
(before) and Prototype 4 (after) applying the guideline # 3.
Prototypes 5 and 6. Tapping outside small feature info to close it appeared to
be challenging for users in comparison to using visible “X” button. On average,
applying the guideline reduced task completion time by 42%, and increased task
completion rate and SUS score by 45 and 22% respectively (Figure 16). The guideline
# 4 was found to have significant effect on all three: task completion time (W = 196,
Z = 3.40, p < 0.001, r = 0.54), task completion rate ( (1, N = 40) = 9.09, p < 0.005),
and SUS score (W = 26.5, Z = -2.58, p = 0.01, r = 0.41).
Figure 16. Task completion time and rate, and SUS scores for Prototype 6
(before) and Prototype 5 (after) applying the guideline # 4.
48
Prototypes 7 and 8. It was natural for users to close feature info using Android
system’s “Back” button. Thus, introducing “Back to the map” button WITHIN
fullscreen feature info yielded smaller effect than described in the previous paragraph.
On average, applying the guideline reduced task completion time by 31%, and
increased task completion rate and SUS score by 5% and 13% respectively (Figure
17). The guideline # 4 was found to have significant effect on task completion time
(W = 180, Z = 2.8, p < 0.005, r = 0.44), task completion rate ( (1, N = 40) = 12.19,
p < 0.001), and SUS score (W = 36.5, Z = -2.13, p < 0.05, r = 0.38).
Figure 17. Task completion time and rate, and SUS scores for Prototype 8
(before) and Prototype 7 (after) applying the guideline # 4.
Overall, results obtained from testing Prototypes 5, 6, 7, and 8 supported
Hypothesis 2.
Prototypes 9 and 10. Because feature info was small and users could see the
map, it was very easy to point out the highlighted map feature after it was selected.
On average, applying the guideline reduced task completion time by 48%, and
increased task completion rate and SUS score by 65% and 57% respectively (Figure
18). The guideline # 8 was found to have significant effect both on task completion
time (W = 210, Z = 3.91, p < 0.001, r = 0.62) and SUS score (W = 0, Z = -3.86, p <
0.001, r = 0.61). It was also found to have insignificant effect on task completion rate
( (1, N = 40) = 1.09, p > 0.05).
49
Figure 18. Task completion time and rate, and SUS scores for Prototype 9
(before) and Prototype 10 (after) applying the guideline #8.
Prototypes 11 and 12. Because feature info was fullscreen and needed to be
closed before returning to map view, it was more challenging to point out the map
feature that was selected. Sometimes it led to confusion because the feature info
window was already closed, but the marker was still highlighted. One user
commented: “I’d prefer seeing the marker highlighted before the [fullscreen] feature
info window comes up. Maybe highlight the marker first, and then after a short delay
make the [fullscreen] feature info window appear”. On average, applying the
guideline allowed reduced task completion time by 27%, and increased task
completion rate and SUS score by 25% and 16% respectively (Figure 19). The
guideline # 8 was found to have significant effect on task completion time (W = 170,
Z = 2.43, p = 0.01, r = 0.38), task completion rate ( (1, N = 40) = 5.76, p < 0.05),
and SUS score (W = 26, Z = -2.40, p < 0.05, r = 0.38). Another critical comment was:
“Changing only the color of the selected marker may not be obvious [because of lack
of contrast]. Changing color and making it bigger would be much better”.
50
Figure 19. Task completion time and rate, and SUS scores for Prototype 11
(before) and Prototype 12 (after) applying the guideline #8.
Overall, results obtained from testing Prototypes 9, 10, 11, and 12 supported
Hypothesis 3.
The correlation between GIS experience and users’ performance and
perception was identified as weak (all r’s ≤ ±0.5). The functionality of the prototypes
was limited since they were only used as proof of concept. Fidelity may have affected
users’ perception and led to high standard deviation of SUS scores. The chosen
prototyping software (POP 2.0 - Prototyping on Paper app) may have affected users’
performance as it did not support scrolling and users had to tap when they wanted to
scroll. Nevertheless, users succeeded in completing the tasks and excluded this
shortcoming from consideration when evaluating the prototypes.
Summarising usability test results, all three guidelines (#3, #4, and #8)
appeared capable of improving the visualization and usability of feature info
interfaces. Several implications for designing feature infos were drawn. First,
maximized (or fullscreen) window size is recommended for displaying long text and
forms. Second, the salience of the control (or button) to close feature info is an
important factor in its interface design. Third, strong visual feedback proved to be
helpful in establishing a mental connection between a selected map feature and its
open feature info window. Although the application of guidelines in this usability test
51
demonstrated significant positive effect on user performance and perception, it is
worth reminding that the test was conducted in a lab-based environment. There is a
possible risk that unobserved normal performance could yield different results as
suggested by Lazar, Feng, and Hochheiser [42]. Another limitation comes from
participants’ demographic characteristics, i.e. all of them were university students of
the same age group. The results of the evaluation might have been different if other
user groups took part in the usability test.
Regardless of what type of control to open/close feature info is chosen; it
would be useful to demonstrate it and other core functionalities to user during
onboarding. Since feature info discoverability was noted as one of the biggest
usability problems, we suggest that further research prioritizes finding ways of
introducing the tool to users. Future work consists of concept testing the rest of the
proposed guidelines. Although the usability test was conducted using tourist
applications, it is also possible to apply established feature info guidelines to mobile
apps of other domains.
All results of this thesis are summarized in Table 8. Answers provided to both
of the research questions are an important contribution to the literature for two
reasons. First, because they facilitate closing the existing “interaction design” gap in
feature info design literature. Second, because they are intended for improving the
Table 8. Results obtained in this thesis linked to the research questions
Research question
# 1: What usability
problems occur when
users interact with feature
info in mobile tourist
applications?
#2: How solutions
addressing these
problems can be
generalized into usability
guidelines for feature info
design?
Main result 34 usability problems
related mostly to
interaction technique,
placement, layout
9 design guidelines with
7 of them focusing on
interaction design
52
usability of mobile maps and taking full advantage of the mobile technology
considering its limitations. These reasons were the motivation for this master thesis,
seen in Section 1.3. Although some ideas presented explicitly help the con terra app
(i.e. results of the participatory design), it is feasible to apply them to any other mobile
map application provided its feature info faces similar challenges and appropriate
adjustments are considered.
53
8. CONCLUSION
This thesis was dedicated to improving the usability of feature info interfaces
in mobile tourist map applications. Dimensions that affect feature info usability were
established. Existing general recommendations for feature info design from academic
and industry sources were consolidated.
The main contribution was nine feature info design guidelines that were
produced to tackle specific usability problems. The guidelines mostly focused on
interaction design of feature infos, an area that received little attention from research
community. In addition, the guidelines were tailored specifically for mobile devices
considering their technical particularities. The usability problems were detected
during heuristic evaluation of five systematically selected mobile tourist applications.
In addition, a user interview was conducted, where one of those five applications was
further scrutinized. Because certain problems lacked obvious solutions, two usability
experts were asked to share their experience of feature info development for mobile
map-based applications. Their responses, existing solutions in evaluated applications,
and supporting academic literature provided the basis for the guidelines.
Three out of nine guidelines were selected for concept testing and initial
evidence of their potential to solve usability problems was obtained. The concepts
produced a significant positive effect on feature info usability, i.e. reduced time
necessary to complete predefined tasks, increased chances of completing tasks
successfully, and were generally well received by the participants. The remaining six
guidelines should also undergo testing prior to being implemented in map products.
Although the guidelines were derived from and tested on tourist applications, they
could be applied to any other domain, provided they match design goals established
by mobile map developers. Once applied, an additional study could be carried out to
collect feature info usability data from a real-world session where users will interact
with a mobile map application unobserved (e.g. A/B testing). Another area for future
research could be feature info interfaces in mobile 3D maps.
54
REFERENCES
[1] A. Skarlatidou, T. Cheng and M. Haklay, "Guidelines for trust interface design
for public engagement Web GIS," International Journal of Geographical
Information, vol. 27, no. 8, pp. 1668-1687, 2013.
[2] I. Muehlenhaus, "Information Window Design," in Web Cartography: Map
Design for Interactive and Mobile Devices, Boca Raton, FL, CRC Press, 2013,
pp. 42-45.
[3] S. Henning and R. Vogler, "User-Centred Map Applications Through
Participatory Design: Experiences Gained During the ‘YouthMap 5020’
Project," The Cartographic Journal, vol. 53, no. 3, pp. 213-229, 2016.
[4] P. A. Longley, M. Goodchild, D. J. Maguire and D. W. Rhind, The Technology
of Problem Solving, 4th ed., Wiley Publishing, 2015.
[5] ISO 9241-210:2010, Ergonomics of human-system interaction — Part 210:
Human-centred design for interactive systems - Terms and definitions, 2010.
[6] J. Earthy, "Development of the usability maturity model," 30 September 1996.
[Online]. Available:
http://www.processforusability.co.uk/Documents/IN511T1.pdf. [Accessed 10
October 2017].
[7] R. Roth, R. Donohue, C. Sack, T. Wallace and T. Buckingham, "A Process for
Keeping Pace with Evolving Web Mapping Technologies," Cartographic
Perspectives, no. 78, pp. 25-52, 2015.
[8] Design Deck, "Map Pop Up (PSD)," 6 December 2011. [Online]. Available:
http://designdeck.co.uk/a/1241. [Accessed 29 September 2017].
[9] P. Merholz, "Google Maps UI - Some thoughts," 17 February 2005. [Online].
Available: http://www.peterme.com/archives/000454.html. [Accessed 12
January 2018].
[10] "Creating a Custom Popup - Google Maps JavaScript API," Google
Developers, 11 December 2017. [Online]. Available:
55
https://developers.google.com/maps/documentation/javascript/examples/overl
ay-popup. [Accessed 12 January 2018].
[11] ESRI, "View pop-ups," [Online]. Available: http://doc.arcgis.com/en/arcgis-
online/get-started/view-pop-ups.htm. [Accessed 15 9 2017].
[12] Agafonkin, Vladimir, "Leaflet Quick Start Guide," 2017. [Online]. Available:
http://leafletjs.com/examples/quick-start/. [Accessed 30 September 2017].
[13] CARTO, "CARTO Editor - Maps - CARTO Documentation," [Online].
Available: https://carto.com/docs/carto-editor/maps/#infowindows. [Accessed
15 9 2017].
[14] Mapbox, "Attach a popup to a marker instance," [Online]. Available:
https://www.mapbox.com/mapbox-gl-js/example/set-popup/. [Accessed 30
September 2017].
[15] The OpenLayers Dev Team, "Popup," [Online]. Available:
https://openlayers.org/en/latest/examples/popup.html. [Accessed 30 September
2017].
[16] ESRI, "Designing Web Map Pop-ups," 10 July 2012. [Online]. Available:
http://storymaps.esri.com/downloads/designingwebmappopups.pdf. [Accessed
10 July 2017].
[17] D. Schobesberger, "Integrating User and Usability Research in Web-Mapping
Application Design," in Modern Trends in Cartography. Lecture Notes in
Geoinformation and Cartography, J. Brus, A. Vondrakova and V. Vozenilek,
Eds., Springer International Publishing, 2015, pp. 147-158.
[18] K. Zickuhr, "Location-Based Services," Pew Research Center, 12 September
2013. [Online]. Available: http://www.pewinternet.org/2013/09/12/location-
based-services/. [Accessed 5 February 2018].
[19] R. Budiu, "Mobile User Experience: Limitations and Strengths," Nielsen
Norman Group, 19 April 2015. [Online]. Available:
https://www.nngroup.com/articles/mobile-ux/. [Accessed 5 February 2018].
56
[20] L. Kuparinen, "Validation and Extension of the Usability Heuristics for Mobile
Map Applications," in International Conference on Cartography and GIS
(6ICC&GIS), Albena, Bulgaria, 2016.
[21] R. Unrau, M. Ostkamp and C. Kray, "An Approach for Harvesting, Visualizing,
and Analyzing WebGIS Sessions to identify usability issues," in Proceedings
of the ACM SIGCHI Symposium on Engineering Interactive Computing
Systems, Lisbon, Portugal, 2017.
[22] G. Chang and L. Caneday, "Web-Based GIS in Tourism Information Search:
An analysis of the effect of socioeconomic characteristics on perception and
behavior," Society and Leisure, vol. 35, no. 1, pp. 155-17, 2012.
[23] G. Chang and L. Caneday, "Web-based GIS in tourism information search:
Perceptions, tasks, and trip attributes," Tourism Management, vol. 32, no. 6, pp.
1435-1437, 2011.
[24] J. Komarkova, O. Visek and M. Novak, "Heuristic evaluation of usability of
GeoWeb sites," in Web and wireless geographical information systems, Berlin
Heidelberg, Springer Verlag, 2007, pp. 264-278.
[25] J. Nielsen, "10 Usability Heuristics for User Interface Design," Nielsen Norman
Group, 1 January 1995. [Online]. Available:
https://www.nngroup.com/articles/ten-usability-heuristics/. [Accessed 25
October 2017].
[26] D. Egginton, O. Turner and T. Wicks, "Common Usability Issues within Web
GIS for Public Information Provision," in The 22nd GIS Research UK
(GISRUK) Conference, Glasgow, UK, 2014.
[27] A.-M. Nivala, S. Brewster and L. T. Sarjakoski, "Usability Evaluation of Web
Mapping Sites," The Cartographic Journal, vol. 45, no. 2, pp. 129-138, 2008.
[28] T. Zhang, N. Kong and I. Stonebraker, "Usability Metrics of Web-based
Mapping Applications," in Digital Library Federation Forum, Atlanta, GA,
2014.
57
[29] R. E. Roth and A. M. MacEachren, "Geovisual analytics and the science of
interaction: an empirical interaction study," Cartography and Geographic
Information Science, vol. 43, no. 1, pp. 30-54, 2016.
[30] E. Park and J. Ohm, "Factors influencing users’ employment of mobile map
services," Telematics and Informatics, vol. 31, no. 2, pp. 253-265, 2014.
[31] A. Hussain, E. Mkpojiogu and M. Yusof, "Perceived usefulness, perceived ease
of use, and perceived enjoyment as drivers for the user acceptance of interactive
mobile maps," in AIP Conference Proceedings, vol. 1761, F. A. A. Nifa, M. N.
M. Nawi and A. Hussain, Eds., Kedah, Malaysia, AIP Publishing, 2016, p.
020051.
[32] A. Garcia, C. Camacho, M. Bellenzier, M. Pasquali, T. Weber and M. Silveira,
"Data Visualization in Mobile Applications: Investigating a Smart City App,"
in Human-Computer Interaction. Interaction Platforms and Techniques. HCI
2016. Lecture Notes in Computer Science, M. Kurosu, Ed., Cham, Springer,
2016, pp. 285-293.
[33] J. Doyle, M. Bertolotto and D. Wilson, "Multimodal interaction-improving
usability and efficiency in a mobile GIS context," in First International
Conference on Advances in Computer-Human Interaction, Sainte Luce,
Martinique, 2008.
[34] J. Feng and Y. Liu, "Intelligent Context-Aware and Adaptive Interface for
Mobile LBS," Computational Intelligence and Neuroscience, pp. 5-15, 2015.
[35] C. Villamor, D. Willis and L. Wroblewski, "Touch Gesture Reference Guide,"
15 April 2010. [Online]. Available:
https://static.lukew.com/TouchGestureGuide.pdf. [Accessed 7 February 2018
].
[36] Google LLC, "Info Windows - Google Maps JavaScript API - Google
Developers," 11 December 2017. [Online]. Available:
https://developers.google.com/maps/documentation/javascript/infowindows.
[Accessed 7 February 2018].
58
[37] R. Adams, "Decision and stress: cognition and e-accessibility in the information
workplace," Universal Access in the Information Society, vol. 5, no. 4, pp. 363-
379, 2007.
[38] N. Ekroth and J. Lennartsson, "Web-based Multicriteria Decision Analysis and
Visualization for Reinvestments in Power Networks," 2017. [Online].
Available: http://www.diva-
portal.org/smash/get/diva2:1119331/FULLTEXT01.pdf. [Accessed 21
October 2017].
[39] P. Haimes, T. Baba and S. Medley, "Mobile map applications and the
democratisation of hazard information," in SIGGRAPH Asia 2015 Mobile
Graphics and Interactive Applications, Kobe, Japan, 2015.
[40] C. Lamsfus, D. Wang, A. Alzua-Sorzabal and Z. Xiang, "Going mobile:
Defining context for on-the-go travelers," Journal of Travel Research, vol. 54,
no. 6, pp. 691-701, 2015.
[41] L. Meng, "The State of the Art of Map-Based Mobile Services," in Map-based
Mobile Services. Lecture Notes in Geoinformation and Cartography, L. Meng,
A. Zipf and S. Winter, Eds., Berlin, Heidelberg, Springer, 2008, pp. 1-12.
[42] J. Lazar, J. Feng and H. Hochheiser, Research Methods in Human-Computer
Interaction, 2nd ed., Morgan Kaufmann, 2017, pp. 39-40.
[43] K. van Turnhout, A. Bennis, S. Craenmehr, R. Holwerda, M. Jacobs, R. Niels,
L. Zaad, S. Hoppenbrouwers, D. Lenior and R. Bakker, "Design patterns for
mixed-method research in HCI," in Proceedings of the 8th Nordic Conference
on Human-Computer Interaction: Fun, Fast, Foundational (NordiCHI '14),
New York, NY, USA, 2014.
[44] J. Nielsen and R. Molich, "Heuristic evaluation of user interfaces," in
Proceedings of the SIGCHI conference on Human factors in computing
systems, Seattle, 1990.
[45] J. Nielsen, "Finding usability problems through heuristic evaluation," in
Proceedings of the SIGCHI conference on Human factors in computing
systems, Monterey, 1992.
59
[46] M. D. Myers, "Qualitative research in information systems," Management
Information Systems Quarterly, vol. 21, no. 2, pp. 241-242, 1997.
[47] J. Webster and R. T. Watson, "Analyzing the past to prepare for the future:
Writing a literature review," MIS quarterly, vol. 26, no. 2, pp. xiii-xxiii, 2002.
[48] G. Gable, "Integrating case study and survey research methods: an example in
information systems," European journal of information systems, vol. 3, no. 2,
pp. 112-126, 1994.
[49] J. Brooke, "SUS: A ‘quick and dirty’ usability scale," in Usability Evaluation
in Industry, P. W. Jordan, B. Thomas, B. A. Weerdmeester and I. L.
McClelland, Eds., London, Taylor & Francis, 1996, pp. 189-194.
[50] J. Sauro and J. R. Lewis, "Standardized Usability Questionnaires," in
Quantifying the user experience: Practical statistics for user research,
Cambridge, MA, Morgan Kaufmann, 2016, pp. 185-186.
[51] T. Tullis and B. Albert, "Measuring the User Experience," 25 September 2017.
[Online]. Available: http://www.measuringux.com/SUS_Calculation.xls.
[Accessed 28 December 2017].
[52] J. Nielsen, "Severity Ratings for Usability Problems," Nielsen Norman Group,
1 January 1995. [Online]. Available: https://www.nngroup.com/articles/how-
to-rate-the-severity-of-usability-problems/. [Accessed 26 09 2017].
[53] Google LLC, "Cards - Components - Material Design," [Online]. Available:
https://material.io/guidelines/components/cards.html. [Accessed 10 November
2017].
[54] A. Karlson, B. Bederson and J. SanGiovanni, "AppLens and launchTile: two
designs for one-handed thumb use on small devices," in Proceedings of the
SIGCHI conference on Human factors in computing systems, Portland, Oregon,
USA, 2005.
[55] N. Andrienko, G. Andrienko, H. Voss, F. Bernardo, J. Hipolito and U.
Kretchmer, "Testing the Usability of Interactive Maps in CommonGIS,"
Cartography and Geographic Information Science, vol. 29, no. 4, pp. 325-342,
2002.
60
[56] "Top Apps Charts," Apple Inc., 2017. [Online]. Available:
https://itunes.apple.com/story/id1297105905. [Accessed 15 December 2017].
[57] "Bleibe zielsicher auf Kurs – zuhause und weltweit," Google Play, 19 October
2017. [Online]. Available:
https://play.google.com/store/apps/topic?id=editorial_navigation_and_maps_d
e. [Accessed 15 December 2017].
[58] J. Lewis and J. Sauro, "The factor structure of the System Usability Scale," in
Human Centered Design, HCII 2009, M. Kurosu, Ed., Heidelberg, Germany,
Springer-Verlag, 2009, pp. 94-103.
[59] T. Hothorn, K. Hornik, M. A. van de Wiel, H. Winell and A. Zeileis, "coin:
Conditional Inference Procedures in a Permutation Test Framework," 28
November 2017. [Online]. Available: https://cran.r-project.org/package=coin.
[Accessed 15 January 2018].
[60] W. Visser, The Cognitive Artifacts of Designing, Mahwah, NJ, USA: Lawrence
Erlbaum Associates, 2006.
61
APPENDIX
A. List of usability problems detected during heuristic evaluation
Con terra app Mobile Google Maps
website
TripAdvisor Google Maps Foursquare
Visibility of system
status.
Text can’t be seen fully,
because feature name takes
up the majority of space
(Severity: 4)
Match between system
and the real world.
Ambiguous attribute names
(Severity: 4)
Use of terminology (e.g.
"GreenLeader Bronze")
without provided
explanation
(Severity: 2)
User control and
freedom.
“Close” button is too small
and touch unfriendly
(Severity: 3)
When accidentally pressing
the "Review" button, there
is no way to discard a
review, only submit
(Severity: 2)
When trying to continue
reading “About” section,
the app takes you to
“Describe this place”
window
(Severity: 4)
"Maximized" windows
covers the map and offers
less space to view feature
info than the "minimized"
one (Severity: 3)
Tiny scrolling area in
feature info allows user to
view only one line of text at
a time (Severity: 4)
When feature info window
is maximized, it’s
impossible to scroll down
because the header takes up
all available space
(Severity: 4)
62
Consistency and
standards.
Mix of German and English
words (Severity: 4)
“Nearest” station is not
available for all landmarks.
Only for those around
Cologne Central station
(Severity: 3)
Feature Info is not available
for all POIs (no feedback
upon touch) (Severity: 4)
Misspelling (Severity: 3)
Missing spaces or dashes
between words
(Severity: 3)
Error prevention. Directions button does
not work at all
(Severity: 4)
Recognition rather than
recall.
When selecting a feature, it
is impossible to distinguish
between the selected and
the rest of the features.
User has to remember what
feature was selected when
going from feature info
dialogue back to the map
(Severity: 4)
Flexibility and efficiency
of use.
Context is ambiguous:
some polygons have an "%
of the total area" field - is
total area the precinct or
city? (Severity: 3)
It is possible to get lost
within "About" map view
(you are presented with a
map and a minimized
feature info, once feature
info is clicked, you return
into the same position you
were previously at. But to
return to the general map
with all features on it, the
user will have to click the
“back” button three times
(Severity: 3)
63
The menu in the bottom
right corner of feature info
is somewhat hidden. It’s
not clear what "1 of 3"
means: pages / features?
(Severity: 3)
Same problem (as
described above) with the
"Address" field
(Severity: 3)
Aesthetic and minimalist
design.
Very little information
presented
(Severity: 3)
“Nearest” station – you
expect to see one station,
but instead you have a
“forward” button that
shows you a list comprised
of only one station in a
strange way
(Severity: 3)
When in the “Photos” tab,
"Videos" are among one of
the subtabs. Confusing
name.
(Severity: 2)
Heading has very large
margins – wasted space
(Severity: 4)
Different visual design of
feature info for every
operational layer
(Severity: 0)
Slightly different design of
feature info for different
operational layers
(Severity: 0)
Feature name is broken into
multiple lines because
doesn’t fit the text box size
(Severity: 4)
Help users recognize,
diagnose, and recover
from errors
Missing information in
some fields and no
indication that it is
missing/when it will be
added
(Severity: 4)
When no feature info for a
POI is available, a feature
info for a previously viewed
POI is shown (Severity: 4)
Missing information about
some POIs (no About field)
(Severity: 2)
Help and documentation.
Total number of errors
found
17 1 8 3 1
Major or catastrophic
problems:
16 1 4 1 1
64
B. Level of detail and actions offered in feature info in evaluated
mobile applications
Maps
Mobile version
of
Google Maps
website
TripAdvisor Foursquare con terra
app
Level of detail
Address Yes Yes Yes Yes Yes
Open hours Yes Yes Yes
Type Yes Yes Yes Yes Yes
Photos Yes Yes Yes Yes
Videos Yes Yes
Review highlights Yes Yes Yes Yes
Review summary Yes Yes Yes Yes
Reviews Yes Yes Yes Yes
Rank Yes Yes Yes Yes
Rating categories Yes Yes
Popular times Yes Yes
Phone Yes Yes Yes Yes
Email Yes
Website Yes Yes Yes Yes Yes
Social media Yes
Similar place suggestions Yes Yes Yes
Nearest station Yes
About Yes Yes Yes
Suggested duration Yes Yes
Questions and answers Yes Yes
Recently viewed Yes
Accessibility Yes
Amenities Yes
Crowd Yes
Integration with other
websites Yes Yes
Upcoming events Yes
65
Features within the currently
viewed Yes
Actions
Directions Yes Yes Yes Yes
Call Yes Yes Yes Yes
Leave a review Yes Yes Yes
Add more information Yes Yes Yes
Add to a list Yes Yes Yes Yes
Check in Yes
Rate Yes Yes Yes
Add photo Yes Yes Yes
Filter photos Yes Yes Yes
Sort photos Yes
Upvote photo Yes
Add tag to photo Yes
Search reviews Yes
Translate reviews Yes Yes
Filter reviews Yes Yes
Sort reviews Yes
Share reviews Yes
Upvote reviews Yes Yes Yes Yes
Make reservations Yes
Ask question Yes Yes
Answer question Yes Yes
Upvote question Yes Yes
Share Yes Yes
Save to favorites Yes
Total 35 17 36 30 3
66
C. User questionnaire
Please answer the questions below in the corresponding answer boxes.
1. Feature Info is a dialogue window that displays background information
about any selected POI. How can feature info help when retrieving additional
background information about single geographic entities?
2. List any issues that you encountered while working with the feature info
window.
3. How did you like the design of the feature info window interface?
4. How did you like that the background information was presented as a table?
5. How would you describe your experience using the feature info window
interface? What aspects appealed to you the most and the least?
6. Why do you think feature info windows should have maximized and
minimized options?
7. How often did you need the map in this application?
8. What will you choose: viewing all markers on the map or having a list of
objects sorted based on proximity to you? Please explain your choice.
9. What information was missing in the feature info windows you browsed?
10. How does the app provide context for the information presented in the feature
info window?
67
11. Why would you want to view background information in an interactive way?
E.g. as infographics. Feel free to list your own examples.
12. When was the last time the background information was updated?
68
D. Expert questionnaire
Please answer the questions below in the corresponding answer boxes.
1. When designing mobile map-based apps, do you implement user-centered
design approach? How?
2. When designing mobile map apps for your clients, do you create custom
feature info windows based on their needs? Are they custom only in content
presented or also in visual design?
Part 1: Visual design
3. What kind of issues do you face when designing a feature info for mobile
apps? E.g. screen real estate in mobile devices
4. Do you think users should be given freedom to customize the look and/or
content of the feature info window? Why?
5. Have you faced situations when map is better and when feature list is better?
6. How do you judge the optimal settings for the following design elements of
the feature info for mobile devices (smartphones):
a. Shape
b. Shape dimensions (size or percentage of the map covered)
c. Position (top, bottom, next to the selected feature)?
7. Is it always necessary to have a full-screen (maximized) mode for the feature
info?
8. How do you emphasize the feature info to attract users’ attention?
69
Part 2: Content
You may answer the questions below in relation to your most recent project that
involved the design of feature info.
9. How do you identify minimum necessary background information to be
displayed in the feature info window?
10. How much functionality should feature info window have?
11. What content types other than text, hyperlinks, and image files do you
incorporate into feature info?
12. Do you think textual information should be presented in the form of table
(see example below) or narrative? Please explain your opinion.
e.g. City: Moscow
Population: 10 million
13. Why do you think providing context in the feature info is important? (e.g.
min and max values, time reference – “closed now”, distance to the POI
from user’s current location).
14. Is there any specific hierarchy in which content should be presented?
Part 3: Usability
15. Have you ever conducted any usability studies for feature info window
interface in mobile map-based applications? Please describe most common
problems / insights of these studies.
16. What issues can users encounter when they are trying to view background
information in mobile map-based applications?
70
E. Tasks designed for prototype usability test
Prototype Task description
1. You want to call the LWL museum to ask about current exhibition.
Please use the app to find the telephone number of LWL museum.
2. You are thinking about visiting the LWL museum but would like to
check it’s rating first. Please use the app to find the review score.
3. You and your friend want to go to Cineplex movie theater and watch
Jumanji: Welcome to the Jungle at 17:00. Please use the app to
reserve tickets.
4. You and your friends want to go to Cineplex movie theater and watch
Jumanji: Welcome to the Jungle at 23:00. Please use the app to
reserve tickets.
5. You want to go to a wine bar. There’s only one bar open right now.
Is it a wine bar? If not, close the window.
6. You want to go out for beers. There’s only one bar open right now.
Do they serve craft beer? If not, close the window.
7. You want to have dinner at a Korean restaurant. There’s only one
restaurant open right now. Do they serve Korean cuisine? If not, close
the window.
8. You want to have dinner at a Russian restaurant. There’s only one
restaurant open right now. Do they serve Russian cuisine? If not,
close the window.
9. You want to go to a night club. Pick one of the clubs on the map and
check it out. After you done, please look back at the map. Can you
point to the exact club that you picked?
10. You want to go to a night club. Pick one of the clubs on the map and
check it out. After you done, please look back at the map. Can you
point to the exact club that you picked?
11. You are located at Komodienstrasse. Browse souvenir shops and pick
one that is located on the same street. After that please return back to
the map. Can you point to the exact shop that you picked?
12. You are located at AlterMarkt. Browse souvenir shops and pick one
that is located on the same street. After that please return back to the
map. Can you point to the exact shop that you picked?
71
F. Declaration of Academic Integrity
I hereby confirm that this thesis on “Feature Info – Improving the visualization and
usability of GIS background information in the context of a mobile tourist
application” is solely my own work and that I have used no sources or aids other than
the ones stated.
All passages in my thesis for which other sources, including electronic media, have
been used, be it direct quotes or content references, have been acknowledged as such
and the sources cited.
13.02.2018,
___________________________________________
(date and signature of student)
I agree to have my thesis checked in order to rule out potential similarities with other
works and to have my thesis stored in a database for this purpose.
13.02.2018,
___________________________________________
(date and signature of student)